1. Field of the Invention
The invention is directed to an improved method of fabricating alloys composed of elements (or other alloys) that have very different melting temperatures and whose components strongly react chemically with each other, especially alloys whose proper function requires strict stoichiometry. The invention is more particularly directed to an improved method of fabricating shape memory alloys.
2. Background of the Invention
The making of alloys is an ancient art. Bronze is stronger than copper alone as a result of alloying copper with zinc, tin, and other elements. Steel is a significant improvement over iron alone, and exhibits greater strength and toughness. Alloys are developed for specific requirements such as resistance to corrosion. Certain alloys exhibiting shape memory properties and superelasticity are used in aerospace, consumer products, and medical devices.
Shape memory alloys (SMAs) are intermetallic compounds that undergo an energetic phase change such that the mechanical properties are very different at temperatures above and below the phase transformation temperature. The most comment SMAs in practical use are Ti-Ni alloys, Cu-based alloys and Fe-based alloys. It has been acknowledged that there are problems with the fabrication of commercial SMAs. In particular, shifting the content of a component, such as Nickel, can result in a change in the martensite start temperature or the transformation finish temperature. See, Shape Memory Materials, J. Otsuka and C. M. Wayman (Eds.), Oxford University Press (1999).
The higher temperature phase, generally referred to as austenite, has a simpler crystal structure and is more rigid than the low temperature phase, called martensite. A body that is deformed (stretched or bent) while in the low temperature phase will recover its original shape when heated to above its transformation temperature, giving rise to a shape memory. In shape memory alloys, the phase transformation from austenite to martensite can be induced by stress. When the stress is removed, the body reverts to its original austenite and consequently recovers its shape. This phenomenon is called superelasticity.
The strain recovery is much greater in shape memory alloys than in ordinary metals. Single crystal SMA may recover as much as the theoretical maximum. For single crystal CuAlNi this is nearly 10 percent strain, which can be described as ‘hyperelastic.’ Thus, while shape memory alloys transform from one solid crystal structure to another, and are capable of energy storage at greater densities than elastic materials, in hyperelastic transformations, the energy is absorbed and released at nearly constant force, so that constant acceleration is attainable. See also, U.S. Patent Publ. 2006/0118210 to Johnson for Portable energy storage devices and methods.
Useful devices are produced by pulling single crystals of CuAlNi from melt by a method due to Vasily Stepanov. Successful pulling of single crystals requires very pure alloys with very strict composition control.
Ingots for crystal pulling are conventionally made by mixing copper, aluminum, and nickel pellets and heating in a furnace. Conventionally, similarly sized pellets of material comprising the alloy are weighed to the fraction that the material represents in the alloy. The pellets are then mixed together in a crucible and heated until the pellets melt and go into solution. The components then engage in congruent melting. FIG. 1 illustrates a CuAlNi ingot 100 cast using conventional methods. This method has several drawbacks. Since the melting temperatures of the elements are disparate, the individual elements comprising the alloy do not readily mix. As the aluminum is melted in conventional methods (ca 650° C.), it becomes reactive and reacts explosively and exothermically with the copper and nickel. This causes spattering, especially in small ingots, and may result in significant loss of mass during the manufacturing process. Additionally, pellets have a large surface to volume ratio and have oxide surfaces that generate slag. Unless the mixture is stirred mechanically, the components can segregate into layers, as illustrated in FIG. 1. Even if stirred vigorously while melted, it may separate during cooling unless cooling is rapid. If a large amount of alloy (more than a kilogram for example) is melted at a time, some segregation is almost certain to occur because cooling cannot be rapid. While the uneven qualities of the ingot of FIG. 1 are an extreme example for an ingot manufactured using conventional methods, significant segregation of components is not unusual. Segregation leads to variation in composition throughout the ingot that causes difficulty in pulling single crystals. Up to 80% of a batch of ingots have been found to be unusable for pulling single crystal shape memory alloy. Even when single crystals are successfully pulled, the results are not reproducible. The transition temperature of the phase transformation that gives CuAlNi its desirable shape memory properties depends crucially on composition (to 0.1 percent).
A method of overcoming these difficulties is important to the commercial development of hyperelastic alloys. The invention described herein is a method whereby small ingots of CuAlNi can be made with reproducible composition and good crystal growth characteristics.